Tool life prediction method, device, equipment and medium based on multi-source data

Abstract

The application relates to the technical field of machining, and discloses a tool life prediction method based on multi-source data: real-time data is acquired, and a tool program is parsed to determine theoretical data; the real-time data is normalized by using the theoretical data to obtain parameter normalization values; all the parameter normalization values and material hardness coefficients are weighted by using a preset fusion model to obtain tool consumption data; a cutting deviation rate and a rotating speed deviation rate are determined, and it is judged whether the preset correction condition is met; when the preset correction condition is met, the tool consumption data is corrected by using the cutting deviation rate and the rotating speed deviation rate to obtain corrected consumption data; the corrected consumption data is subjected to life prediction by using a life prediction model to obtain tool predicted life. The application triggers correction by using the deviation rate, avoids the limitation of fixed weight, simplifies the prediction process by using the life prediction model and the multi-source data, and thus the prediction efficiency and accuracy are improved.

Description

Technical Field

[0001] This invention relates to the field of machining technology, and in particular to a method, apparatus, equipment and medium for predicting tool life based on multi-source data. Background Technology

[0002] During machining, identifying and intelligently detecting the wear condition of cutting tools in CNC machine tools is crucial to ensuring machining accuracy. Different degrees of tool wear have varying impacts on machining efficiency, affecting both the quality and cost of the produced workpieces and negatively impacting production efficiency.

[0003] Existing methods for inspecting CNC machine tool tools require halting production, impacting efficiency. Alternatively, finite element analysis-based tool life prediction methods rely on complex models, resulting in high computational costs and inconsistent accuracy. Other methods use fixed standard curves for tool inspection, but tool wear varies with machining processes and may not perfectly align with these curves, leading to low accuracy. Therefore, a tool life prediction solution is urgently needed to address these technical challenges. Summary of the Invention

[0004] This invention provides a tool life prediction method, device, and medium based on multi-source data to solve the problem of low accuracy in existing tool life prediction methods.

[0005] A tool life prediction method based on multi-source data includes: The system acquires real-time data from multiple sensors and analyzes the tool program sheet to determine theoretical data; the real-time data includes the actual cutting length and the actual spindle speed. The real-time data is normalized using the theoretical data to obtain parameter normalization values ​​corresponding to each of the real-time data. The material hardness coefficient is obtained, and the normalized values ​​of all the parameters and the material hardness coefficient are weighted by a preset fusion model to obtain tool consumption data. Determine the cutting deviation rate corresponding to the actual cutting length and the rotational speed deviation rate corresponding to the actual spindle speed, and determine whether the cutting deviation rate and the rotational speed deviation rate meet the preset correction conditions; When the cutting deviation rate and / or the rotational speed deviation rate meet the preset correction conditions, the tool consumption data is corrected using the cutting deviation rate and the rotational speed deviation rate to obtain corrected consumption data; Obtain the lifespan preset model, and use the lifespan prediction model to predict the lifespan of the corrected consumption data to obtain the tool prediction lifespan.

[0006] In one embodiment, the theoretical data includes theoretical cutting length and theoretical spindle speed; The determination of the cutting deviation rate corresponding to the actual cutting length and the speed deviation rate corresponding to the actual spindle speed includes: The absolute value of the cutting difference between the actual cutting length and the theoretical cutting length is determined, and the cutting deviation rate is determined based on the absolute value of the cutting and the theoretical cutting length. Determine the absolute value of the difference between the actual spindle speed and the theoretical spindle speed, and determine the speed deviation rate based on the absolute value of the speed and the theoretical spindle speed.

[0007] In one embodiment, the preset correction conditions include a preset cutting threshold and a preset rotation speed threshold; The step of determining whether the cutting deviation rate and the rotational speed deviation rate meet the preset correction conditions includes: The cutting deviation rate and the preset cutting threshold are compared to obtain the cutting comparison result. The speed deviation rate and the preset speed threshold are compared to obtain the speed comparison result. When the cutting comparison result indicates that the cutting deviation rate is greater than the preset cutting threshold, and / or the rotational speed comparison result indicates that the rotational speed deviation rate is greater than the preset rotational speed threshold, it is determined that the cutting deviation rate and the rotational speed deviation rate meet the preset correction condition; When the cutting comparison result indicates that the cutting deviation rate is less than or equal to the preset cutting threshold, and the rotational speed comparison result indicates that the rotational speed deviation rate is less than or equal to the preset rotational speed threshold, it is determined that the cutting deviation rate and the rotational speed deviation rate do not meet the preset correction conditions.

[0008] The process of correcting the tool consumption data using the cutting deviation rate and the rotational speed deviation rate to obtain corrected consumption data includes: Obtain a preset correction coefficient, and determine a correction deviation value based on the preset correction coefficient, the cutting deviation rate, and the rotational speed deviation rate; The tool consumption data is corrected by using the correction deviation value to obtain corrected consumption data.

[0009] In one embodiment, the preset fusion model includes: Life_Index=C1T norm +C2L norm +C3S norm +C4F norm +C5M; in: Life_Index represents tool consumption data; C1, C2, C3, C4, and C5 are preset weighting coefficients; T norm This is a time normalization value; L norm This is the normalized value for cutting. S norm This is the normalized value for rotational speed; F norm This is the load normalization value; M is the material hardness coefficient of the cutting tool.

[0010] In one embodiment, after weighting all the normalized values ​​of the parameters and the material hardness coefficient using a preset fusion model to obtain the tool consumption data, the method further includes: Obtain all warning levels, and associate each warning level with a warning index range; According to the order of the warning levels from high to low, the tool consumption data is compared with the warning index range corresponding to each warning level in turn to determine the warning index range to which the tool consumption data belongs; The warning level corresponding to the warning index range to which the tool consumption data belongs is determined as the target warning level, and the triggering of the target warning level is determined.

[0011] In one embodiment, before obtaining the material hardness coefficient, the method further includes: Obtain a sample dataset, which includes at least one set of sample normalization values ​​and a sample consumption index corresponding to each set of sample normalization values; A preset training model is obtained, and the normalized values ​​of all the samples are weighted using the preset training model to obtain the predicted consumption index. The prediction loss value of the preset training model is determined based on the sample consumption index and the prediction consumption index corresponding to the same sample normalization value. When the predicted loss value does not reach the preset convergence condition, the initial parameters in the preset training model are iteratively updated until the predicted loss value reaches the convergence condition. Then, the converged preset training model is recorded as the preset fusion model.

[0012] A tool life prediction device based on multi-source data, comprising: The data acquisition module is used to acquire real-time data measured by multiple sensors and to analyze the tool program sheet to determine theoretical data; the real-time data includes the actual cutting length and the actual spindle speed. The data normalization module is used to normalize the real-time data using the theoretical data to obtain parameter normalization values ​​corresponding to each of the real-time data. The consumption index module is used to obtain the material hardness coefficient. By using a preset fusion model, the normalized values ​​of all the parameters and the material hardness coefficient are weighted to obtain tool consumption data. The deviation rate determination module is used to determine the cutting deviation rate corresponding to the actual cutting length and the rotational speed deviation rate corresponding to the actual spindle speed, and to determine whether the cutting deviation rate and the rotational speed deviation rate meet the preset correction conditions. The data correction module is used to correct the tool consumption data by means of the cutting deviation rate and / or the speed deviation rate when the cutting deviation rate and / or the speed deviation rate meet the preset correction conditions, so as to obtain corrected consumption data. The life prediction module is used to acquire a life preset model, and to predict the life of the tool by using the life prediction model to predict the life of the corrected consumption data.

[0013] A computer device includes a memory, a processor, and a computer program stored in the memory and executable on the processor, the processor being used to perform the above-described tool life prediction method based on multi-source data.

[0014] A computer-readable storage medium storing a computer program that, when executed by a processor, implements the above-described tool life prediction method based on multi-source data.

[0015] The aforementioned tool life prediction method, apparatus, equipment, and medium based on multi-source data, particularly the tool life prediction method based on multi-source data of this invention, avoids the problems of offline prediction and online failure by utilizing real-time data and theoretical data from the tool program sheet. By pre-setting a fusion model, cutting deviation rate, and speed deviation rate, tool consumption data is calculated and subsequently corrected, avoiding the limitations of fixed weights and enabling the calculation of tool consumption data based on machining conditions. Through the life prediction model and multi-source data, the remaining tool life is predicted, simplifying the prediction process and thus improving prediction efficiency and accuracy. Attached Figure Description

[0016] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the description of the embodiments of the present invention will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0017] Figure 1 This is a flowchart of a tool life prediction method based on multi-source data in one embodiment of the present invention; Figure 2This is a schematic diagram of a tool life prediction device based on multi-source data in one embodiment of the present invention. Detailed Implementation

[0018] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the protection scope of the present invention.

[0019] In one embodiment, such as Figure 1 As shown, a tool life prediction method based on multi-source data is provided, including the following steps: S10: Acquire real-time data measured by multiple sensors and parse the tool program sheet to determine theoretical data; the real-time data includes the actual cutting length and the actual spindle speed.

[0020] Understandably, a tool program is used to record parameters for each machining step in CNC machining. Real-time data refers to the machining parameters of the CNC machine tool measured by sensors. Actual cutting length refers to the actual length of the workpiece cut by the tool during the cutting process, such as 300 meters, 800 meters, etc. Actual spindle speed refers to the current spindle speed of the CNC machine tool. Theoretical data refers to data derived from tool parameters through mathematical derivation, logical reasoning, or computer simulation.

[0021] Specifically, by installing multiple different types of sensors at different locations on the CNC machine tool, real-time data of the CNC machine tool under the current time period and operating conditions is acquired, such as spindle load, actual cutting length, and actual spindle speed. Then, tool programs are selected from the bill of materials data, and the data in the tool programs is parsed to obtain theoretical data, such as theoretical cutting time, theoretical cutting length, and theoretical speed.

[0022] S20: Normalize the real-time data using the theoretical data to obtain parameter normalization values ​​corresponding to each of the real-time data.

[0023] Understandably, the parameter normalization value refers to the value after normalization for each real-time data, such as parameters like spindle speed, cutting length, and spindle load.

[0024] Specifically, real-time data is normalized using theoretical data. This involves obtaining a preset normalization model, inputting all real-time and theoretical data into this model, and then normalizing the real-time data and its corresponding theoretical data using the model based on a learned normalization formula. This yields the normalized parameter values ​​for each real-time data point. This process is repeated for all real-time data to obtain the corresponding parameter normalization values. = , For parameter normalization value, For real-time data, These are theoretical data. For example, normalized speed values, normalized load values, etc.

[0025] S30: Obtain the material hardness coefficient, and perform weighted processing on the normalized values ​​of all the parameters and the material hardness coefficient through a preset fusion model to obtain tool consumption data.

[0026] Understandably, a pre-defined fusion model refers to a neural network model trained on a large amount of sample data, possessing weights corresponding to each real-time data point. Tool consumption data is used to measure the degree of tool wear during machining.

[0027] Specifically, a preset fusion model is obtained, and the normalized values ​​of all parameters are input into the preset fusion model. Then, the material hardness coefficient corresponding to the tool is obtained and input into the preset fusion model. The preset fusion model performs a weighted fusion of all normalized parameter values ​​and material hardness coefficients. That is, first, the weights corresponding to the normalized values ​​of each parameter and the weights corresponding to the material hardness coefficients, which were learned by the preset fusion model during training, are obtained. Then, the product of the weights and the normalized parameter values, and the product of the material hardness coefficients and the weights are calculated. Finally, all the calculated product results are added together to obtain the tool consumption data.

[0028] S40: Determine the cutting deviation rate corresponding to the actual cutting length and the rotational speed deviation rate corresponding to the actual spindle speed, and determine whether the cutting deviation rate and the rotational speed deviation rate meet the preset correction conditions.

[0029] Understandably, preset correction conditions refer to the conditions used to evaluate whether tool consumption data has been corrected. Cutting deviation rate refers to the deviation rate between the actual cutting length and the theoretical cutting length. Spindle speed deviation rate refers to the deviation between the actual spindle speed and the theoretical spindle speed.

[0030] Specifically, the real-time cutting length and theoretical cutting length are obtained. Then, the cutting deviation rate is calculated based on these two lengths; that is, the difference between the calculated real-time cutting length and the theoretical cutting length is divided by the theoretical cutting length. Similarly, the real-time spindle speed and theoretical spindle speed are obtained. Then, the speed deviation rate is calculated based on these two speeds; that is, the difference between the calculated real-time spindle speed and the theoretical spindle speed is divided by the theoretical spindle speed. Next, preset correction conditions are obtained, and the cutting deviation rate and speed deviation rate are compared with these conditions. If at least one of the cutting deviation rate and speed deviation rate meets the preset correction condition, it is determined that the cutting deviation rate and speed deviation rate meet the correction condition. If neither the cutting deviation rate nor the speed deviation rate meets the preset correction condition, it is determined that neither the cutting deviation rate nor the speed deviation rate meets the correction condition.

[0031] S50: When the cutting deviation rate and / or the rotational speed deviation rate meet the preset correction conditions, the tool consumption data is corrected by the cutting deviation rate and the rotational speed deviation rate to obtain corrected consumption data.

[0032] Understandably, the corrected consumption data refers to the actual tool consumption data after correction by the cutting deviation rate and the speed deviation rate.

[0033] Specifically, when the cutting deviation rate and / or speed deviation rate meet the preset correction conditions, the tool consumption data is corrected by the cutting deviation rate and speed deviation rate. That is, the preset correction coefficient is obtained, the correction deviation value is calculated first based on the preset correction coefficient, cutting deviation rate and speed deviation rate, and then the data correction operation is performed on the tool consumption data based on the correction deviation value. That is, the sum of the correction deviation value and the tool consumption data is calculated, and the result is determined as the correction consumption data.

[0034] S60: Obtain the life preset model, and use the life prediction model to predict the life of the corrected consumption data to obtain the tool prediction life.

[0035] In essence, tool life prediction refers to the predicted remaining tool life. A tool life prediction model is a model used to predict tool life, which is trained based on a large amount of sample data.

[0036] Specifically, a tool life prediction model is obtained, and the corrected consumption data is input into the model. The model then predicts the tool life based on the corrected consumption data. Specifically, the tool consumption rate is first calculated based on tool consumption data from different time periods, i.e., the historical consumption index calculated in the previous time period is obtained. The tool consumption rate is then calculated based on the current tool consumption data and the historical consumption index. Finally, the predicted tool life is calculated based on the tool consumption data and the tool consumption rate. This involves first calculating the remaining tool index based on the tool consumption data, and then dividing the remaining tool index by the tool consumption rate to determine the predicted tool life. Where Remaining_Time = , For tool consumption data, Current_Index_Rate represents the tool wear rate, and Remaining_Time represents the predicted tool life. , For tool consumption data, This is a historical consumption index. The unit of time is the time difference between calculating two sets of tool consumption data.

[0037] The tool life prediction method based on multi-source data in this invention avoids the problems of offline prediction and online failure by using real-time data and theoretical data from the tool program sheet. By pre-setting a fusion model, cutting deviation rate, and spindle speed deviation rate, tool consumption data is calculated and corrected, avoiding the limitations of fixed weights and enabling the calculation of tool consumption data based on machining conditions. Through the life prediction model and multi-source data, the remaining tool life is predicted, simplifying the prediction process and improving its efficiency and accuracy.

[0038] In one embodiment, after step S30, that is, after weighting all the normalized values ​​of the parameters and the material hardness coefficient using a preset fusion model to obtain the tool consumption data, the method further includes: S301: Obtain all warning levels, with each warning level associated with a warning index range.

[0039] S302, according to the order of the warning levels from high to low, the tool consumption data is compared with the warning index range corresponding to each warning level in turn to determine the warning index range to which the tool consumption data belongs.

[0040] S303, determine the warning level corresponding to the warning index range to which the tool consumption data belongs as the target warning level, and determine to trigger the target warning level.

[0041] In essence, a warning level refers to a graded indication of tool consumption data usage information, such as a yellow prediction or a red warning. The warning index range refers to the magnitude range of the tool consumption data, for example, 0.5 to 0.6, or 0.8 to 0.9. The target warning level refers to the warning level corresponding to the warning index range in which the tool consumption data falls.

[0042] Specifically, after obtaining the tool consumption data, all warning levels are acquired, and a warning index range is associated with each warning level. Then, the tool consumption data is compared sequentially with the warning index range corresponding to each warning level, either in descending order of warning level or in ascending order of warning level, to determine the warning index range to which the tool consumption data belongs. That is, the tool consumption data is compared with the warning index range of the highest warning level. If the tool consumption data falls within the warning index range of the highest warning level, the highest warning level is triggered and designated as the target warning level. If the tool consumption data does not fall within the warning index range of the highest warning level, the warning index range of the next warning level is compared with the tool consumption data until the warning index range to which the tool consumption data belongs is determined. The warning level corresponding to the warning index range to which the tool consumption data belongs is then designated as the target warning level, and the target warning level is triggered.

[0043] In this embodiment, by using tool consumption data and warning levels, the warning index range to which the tool consumption data belongs is determined, thereby enabling the determination and triggering of the target warning level, and thus avoiding collisions caused by tool wear, which would reduce the yield of processed products.

[0044] In one embodiment, step S40, namely the theoretical data, includes the theoretical cutting length and the theoretical spindle speed; The determination of the cutting deviation rate corresponding to the actual cutting length and the speed deviation rate corresponding to the actual spindle speed includes: S401, determine the absolute value of the difference between the actual cutting length and the theoretical cutting length, and determine the cutting deviation rate based on the absolute value of the cutting length and the theoretical cutting length.

[0045] S402, determine the absolute value of the difference between the actual spindle speed and the theoretical spindle speed, and determine the speed deviation rate based on the absolute value of the speed and the theoretical spindle speed.

[0046] Understandably, the theoretical data includes theoretical cutting length and theoretical spindle speed. The theoretical cutting length refers to the total length of the workpiece cut by the tool under ideal conditions, for example, 3000 meters. The theoretical spindle speed refers to the preset rotational speed of the machine tool spindle. The absolute cutting value refers to the absolute value of the difference between the actual cutting length and the theoretical cutting length. The absolute spindle speed refers to the absolute value of the actual spindle speed and the theoretical spindle speed. It is understood that the calculation of these two deviation values ​​in this embodiment is merely illustrative and not intended to be limiting; time deviation values, feed rate deviation values, etc., can also be added.

[0047] Specifically, after obtaining tool consumption data through a preset fusion model, the theoretical cutting length and theoretical spindle speed are acquired from the theoretical data. Then, the difference between the actual cutting length and the theoretical cutting length is calculated, and the absolute value of this difference is taken as the absolute cutting value. Finally, based on the absolute cutting value and the theoretical cutting length, the cutting deviation rate is determined; that is, the quotient of the absolute cutting value and the theoretical cutting length is determined as the cutting deviation rate. , This refers to the cutting deviation rate. This is the actual cutting length. This represents the theoretical cutting length. Similarly, calculate the difference between the actual spindle speed and the theoretical spindle speed, and take the absolute value of this difference as the absolute value of the spindle speed. Then, based on the absolute value of the spindle speed and the theoretical spindle speed, determine the cutting deviation rate; that is, the quotient of the absolute value of the spindle speed and the theoretical spindle speed is determined as the spindle speed deviation rate. , This refers to the cutting deviation rate. This is the actual cutting length. This is the theoretical cutting length.

[0048] In this embodiment, the absolute value of the cutting length is calculated by using the actual cutting length and the theoretical cutting length, thereby enabling the calculation of the cutting deviation value. Similarly, the cutting speed value is calculated by using the actual spindle speed and the theoretical spindle speed, thereby enabling the calculation of the speed deviation value and improving the accuracy of subsequent tool life prediction.

[0049] In one embodiment, step S40, namely the preset correction conditions, includes a preset cutting threshold and a preset rotation speed threshold. The step of determining whether the cutting deviation rate and the rotational speed deviation rate meet the preset correction conditions includes: S403, compare the cutting deviation rate and the preset cutting threshold to obtain the cutting comparison result.

[0050] S404, compare the speed deviation rate and the preset speed threshold to obtain the speed comparison result.

[0051] S405, when the cutting comparison result indicates that the cutting deviation rate is greater than the preset cutting threshold, and / or the speed comparison result indicates that the speed deviation rate is greater than the preset speed threshold, it is determined that the cutting deviation rate and the speed deviation rate meet the preset correction condition.

[0052] S406, when the cutting comparison result indicates that the cutting deviation rate is less than or equal to the preset cutting threshold, and the rotational speed comparison result indicates that the rotational speed deviation rate is less than or equal to the preset rotational speed threshold, it is determined that the cutting deviation rate and the rotational speed deviation rate do not meet the preset correction condition.

[0053] Understandably, the preset correction conditions include a preset cutting threshold and a preset rotational speed threshold. The preset cutting threshold refers to the cutting threshold at which a pre-set cutting deviation rate triggers tool consumption data correction. The preset rotational speed threshold refers to the rotational speed threshold at which a pre-set rotational speed deviation rate triggers tool consumption data correction. The cutting comparison result is used to characterize the magnitude of the cutting deviation rate and the preset cutting threshold. The rotational speed comparison result is used to characterize the magnitude of the rotational speed deviation rate and the preset rotational speed threshold.

[0054] Specifically, after determining the cutting deviation rate and the speed deviation rate, preset correction conditions are obtained, and preset cutting thresholds and preset speed thresholds are determined within these preset correction conditions. Then, the cutting deviation rate and the preset cutting thresholds are compared to obtain a cutting comparison result indicating that the cutting deviation rate is greater than the preset cutting threshold, or a cutting comparison result indicating that the cutting deviation rate is less than or equal to the preset cutting threshold. Similarly, the speed deviation rate and the preset speed threshold are compared to obtain a cutting comparison result indicating that the speed deviation rate is greater than the preset speed threshold, or a speed comparison result indicating that the speed deviation rate is less than or equal to the preset speed threshold. Further, when the cutting comparison result indicates that the cutting deviation rate is greater than the preset cutting threshold, and / or the speed comparison result indicates that the speed deviation rate is greater than the preset speed threshold, it is determined that the cutting deviation rate and the speed deviation rate satisfy the preset correction conditions; that is, when at least one of the cutting deviation rate and the speed deviation rate is greater than a preset threshold, it is determined that the cutting deviation rate and the speed deviation rate satisfy the preset correction conditions. Next, when the cutting deviation rate, as indicated by the cutting comparison result, is less than or equal to the preset cutting threshold, and the speed deviation rate, as indicated by the speed comparison result, is less than or equal to the preset speed threshold, it is determined that the cutting deviation rate and the speed deviation rate do not meet the preset correction conditions. That is, when both the cutting deviation rate and the speed deviation rate are not greater than the preset threshold, it is determined that neither the cutting deviation rate nor the speed deviation rate meets the preset correction conditions.

[0055] In this embodiment, by setting a cutting threshold and a speed threshold, the cutting deviation value and the speed deviation value are compared, thereby realizing the evaluation of whether the preset correction conditions are met, and thus improving the accuracy of subsequent tool life prediction.

[0056] In one embodiment, step S50, namely, correcting the tool consumption data using the cutting deviation rate and the rotational speed deviation rate to obtain corrected consumption data, includes: S501, obtain a preset correction coefficient, and determine a correction deviation value based on the preset correction coefficient, the cutting deviation rate, and the rotational speed deviation rate.

[0057] S502, the tool consumption data is corrected using the correction deviation value to obtain corrected consumption data.

[0058] Understandably, a preset correction factor refers to a factor determined through analysis and calculation of a large amount of data, used to correct tool consumption data based on the deviation rate. Examples include 0.2, 0.3, etc. The correction deviation value refers to the magnitude of the correction needed to the tool consumption data.

[0059] Specifically, when it is confirmed that the cutting deviation rate and the speed deviation rate meet the preset correction conditions, a preset correction coefficient is obtained. Based on the preset correction coefficient, the cutting deviation rate, and the speed deviation rate, the correction deviation value is determined, that is, a preset calculation model is obtained. The preset correction coefficient, the cutting deviation rate, and the speed deviation rate are input into the preset calculation model. The preset calculation model is then used to calculate the preset correction coefficient, the cutting deviation rate, and the speed deviation rate using the calculation formula learned during training. That is, the sum of the cutting deviation rate and the speed deviation rate is first calculated, and then the product of the sum and the preset correction coefficient is calculated to obtain the correction deviation value. For example, the calculation formula is M= k (Δ L +Δ S ), where M is the correction deviation value, k Δ is the preset correction factor. L For cutting deviation rate, Δ S This represents the rotational speed deviation rate. Further, the tool consumption data is corrected by adjusting the deviation value; that is, the sum of the tool consumption data and the correction deviation value is calculated to obtain the actual tool consumption data, which is then determined as the corrected consumption data.

[0060] In this embodiment, by presetting correction coefficients, cutting deviation rate, and rotational speed deviation rate, the correction deviation value is calculated, thereby realizing the correction of tool consumption data and the acquisition of correction consumption data, which in turn improves the accuracy of tool consumption calculation and the accuracy of tool remaining life prediction.

[0061] In one embodiment, step S30, i.e., the preset fusion model, includes: Life_Index=C1T norm +C2L norm +C3S norm +C4F norm +C5M; in: Life_Index represents tool consumption data; C1, C2, C3, C4, and C5 are preset weighting coefficients; T norm This is a time normalization value; L norm This is the normalized value for cutting. S norm This is the normalized value for rotational speed; F norm This is the load normalization value; M is the material hardness coefficient of the cutting tool.

[0062] In one embodiment, the preset weighting coefficients can be the same or different, depending on the actual situation.

[0063] In one embodiment, before step S30, i.e. before obtaining the material hardness coefficient, the method further includes: S701, Obtain a sample dataset, the sample dataset including at least one set of sample normalization values ​​and a sample consumption index corresponding to each set of sample normalization values.

[0064] In essence, a sample normalization value refers to a parameter normalization value calculated based on historical data, along with the corresponding sample hardness coefficient. The sample hardness coefficient refers to the material hardness coefficient of the cutting tool. Each set of sample normalization values ​​includes the normalization values ​​of multiple parameters, and each set is associated with a sample consumption index. The sample consumption index refers to the tool consumption data calculated based on preset weights and each set of sample normalization values. This data can also be calculated and corrected based on other models. Each set of sample normalization values, sample hardness coefficients, and sample consumption indices can be obtained from data of different historical time periods from different databases, or they can be pre-prepared data sent from the client to the database. A sample dataset is then constructed based on all the obtained sample normalization values, sample hardness coefficients, and sample consumption indices.

[0065] S702, Obtain a preset training model, and use the preset training model to perform weighted processing on the normalized values ​​of all the samples to obtain the predicted consumption index.

[0066] In essence, the predicted consumption index refers to tool consumption information generated by a pre-trained model based on the normalized value and hardness coefficient of each sample. The pre-trained model can be a neural network model or an open-source AI model.

[0067] Specifically, after acquiring the sample dataset, a pre-trained model is invoked, and all sample normalization values, sample hardness coefficients, and sample consumption indices are input into the pre-trained model. This allows the pre-trained model to learn tool consumption calculation capabilities based on all sample normalization values, sample consumption indices, and sample hardness coefficients. In other words, the sample normalization values ​​and sample hardness coefficients are used as model inputs, and the sample consumption index is used as model outputs. This allows the pre-trained model to continuously learn and acquire consumption calculation capabilities. The pre-trained model learns the weights corresponding to each sample normalization value and the weights corresponding to the sample hardness coefficients, enabling it to perform exponential consumption processing based on the sample normalization values ​​and sample hardness coefficients to obtain the predicted consumption index.

[0068] S703, determine the prediction loss value of the preset training model based on the sample consumption index and the prediction consumption index corresponding to the same sample normalization value.

[0069] Understandably, the predicted loss value is generated during the process of calculating the normalized value of the sample by the preset training model.

[0070] Specifically, after obtaining the predicted consumption index, all predicted consumption indices corresponding to the sample normalization values ​​are arranged according to the order of the sample normalization values ​​in the sample dataset. Then, the predicted consumption index associated with the sample normalization value is compared with the sample consumption index of the sample normalization value in the same sequence. That is, according to the sample normalization value, the sample consumption index corresponding to the first sample normalization value is compared with the predicted consumption index corresponding to the first sample normalization value, and the loss value between the sample consumption index and the predicted consumption index is determined by the loss function. The sample consumption index corresponding to the second sample normalization value is compared with the predicted consumption index corresponding to the second sample normalization value, and the loss value between the sample consumption index and the predicted consumption index is determined by the loss function. This process continues until all sample consumption indices and predicted consumption indices have been compared, and the predicted loss value of the preset training model is obtained.

[0071] S704, when the predicted loss value does not reach the preset convergence condition, iteratively update the initial parameters in the preset training model until the predicted loss value reaches the convergence condition, and record the converged preset training model as the preset fusion model.

[0072] Understandably, the preset convergence condition can be either the predicted loss value being less than a set threshold, or the predicted loss value being very small and no longer decreasing after 50,000 calculations, at which point training can be stopped.

[0073] Specifically, when the predicted loss value fails to meet the preset convergence condition, the initial parameters of the preset training model are adjusted based on the predicted loss value. All sample normalization values, sample hardness coefficients, and all sample consumption indices are then re-inputted into the preset training model with adjusted initial parameters. Optimizing the parameters in the preset training model yields the predicted loss value corresponding to the preset training model with adjusted initial parameters. Then, when the predicted loss value fails to meet the preset convergence condition again, the initial parameters of the preset training model are adjusted again based on the predicted loss value, until the predicted loss value of the preset training model with adjusted initial parameters meets the preset convergence condition. In this way, the output of the preset training model continuously approaches the accurate result, resulting in increasingly higher prediction accuracy, until the predicted loss values ​​of all sample normalization values, all sample hardness coefficients, and all sample consumption indices all meet the preset convergence condition. At this point, the converged preset training model is recorded as the preset fusion model.

[0074] In this embodiment, the preset training model is trained by a large number of sample normalization values ​​and sample consumption index, and the overall loss value of the preset training model is calculated by the loss function. This realizes the determination of the predicted loss value of the preset training model, thereby realizing the acquisition of the preset fusion model and ensuring that the preset fusion model has a high accuracy.

[0075] It should be understood that the sequence number of each step in the above embodiments does not imply the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present invention.

[0076] In one embodiment, a tool life prediction device based on multi-source data is provided, which corresponds one-to-one with the tool life prediction method based on multi-source data in the above embodiments. For example... Figure 2 As shown, the tool life prediction device based on multi-source data includes a data acquisition module 10, a data normalization module 20, a consumption index module 30, a deviation rate determination module 40, a data correction module 50, and a tool life prediction module 60. Detailed descriptions of each functional module are as follows: The data acquisition module 10 is used to acquire real-time data measured by multiple sensors and to analyze the tool program sheet to determine theoretical data; the real-time data includes the actual cutting length and the actual spindle speed. The data normalization module 20 is used to normalize the real-time data using the theoretical data to obtain parameter normalization values ​​corresponding to each of the real-time data. The consumption index module 30 is used to obtain the material hardness coefficient. By using a preset fusion model, the normalized values ​​of all the parameters and the material hardness coefficient are weighted to obtain tool consumption data. The deviation rate determination module 40 is used to determine the cutting deviation rate corresponding to the actual cutting length and the rotational speed deviation rate corresponding to the actual spindle speed, and to determine whether the cutting deviation rate and the rotational speed deviation rate meet the preset correction conditions. The data correction module 50 is used to correct the tool consumption data by means of the cutting deviation rate and the speed deviation rate when the cutting deviation rate and / or the speed deviation rate meet the preset correction conditions, so as to obtain corrected consumption data. The life prediction module 60 is used to acquire a life preset model, and to predict the life of the tool by using the life prediction model to predict the life of the corrected consumption data.

[0077] In one embodiment, the theoretical data includes theoretical cutting length and theoretical spindle speed; the deviation rate determination module 40 includes: The cutting deviation rate unit is used to determine the absolute value of the difference between the actual cutting length and the theoretical cutting length, and to determine the cutting deviation rate based on the absolute value of the cutting length and the theoretical cutting length. The speed deviation rate unit is used to determine the absolute value of the difference between the actual spindle speed and the theoretical spindle speed, and to determine the speed deviation rate based on the absolute value of the speed and the theoretical spindle speed.

[0078] In one embodiment, the preset correction conditions include a preset cutting threshold and a preset rotation speed threshold; The deviation rate determination module 40 also includes: A cutting comparison unit is used to compare the cutting deviation rate and the preset cutting threshold to obtain a cutting comparison result. The speed comparison unit is used to compare the speed deviation rate and the preset speed threshold to obtain the speed comparison result. The condition satisfaction unit is used to determine that the cutting deviation rate and the speed deviation rate satisfy a preset correction condition when the cutting comparison result indicates that the cutting deviation rate is greater than the preset cutting threshold, and / or the speed comparison result indicates that the speed deviation rate is greater than the preset speed threshold. The condition-not-met unit is used to determine that the cutting deviation rate and the speed deviation rate do not meet the preset correction conditions when the cutting comparison result indicates that the cutting deviation rate is less than or equal to the preset cutting threshold, and the speed comparison result indicates that the speed deviation rate is less than or equal to the preset speed threshold.

[0079] In one embodiment, the data correction module 50 includes: A deviation correction unit is used to obtain a preset correction coefficient and determine a correction deviation value based on the preset correction coefficient, the cutting deviation rate, and the rotational speed deviation rate. The data correction unit is used to correct the tool consumption data using the correction deviation value to obtain corrected consumption data.

[0080] In one embodiment, the consumption index module 30 includes: Life_Index=C1T norm +C2L norm +C3S norm +C4F norm +C5M; in: Life_Index represents tool consumption data; C1, C2, C3, C4, and C5 are preset weighting coefficients; T norm This is a time normalization value; L norm This is the normalized value for cutting. S norm This is the normalized value for rotational speed; F norm This is the load normalization value; M is the material hardness coefficient of the cutting tool.

[0081] In one embodiment, the device further includes: The warning level unit is used to obtain all warning levels, and each warning level is associated with a warning index range; The warning comparison unit is used to compare the tool consumption data with the warning index range corresponding to each warning level in descending order of the warning level, so as to determine the warning index range to which the tool consumption data belongs. The early warning triggering unit is used to determine the early warning level corresponding to the early warning index range to which the tool consumption data belongs as the target early warning level, and to determine to trigger the target early warning level.

[0082] In one embodiment, the device further includes: A sample acquisition unit is used to acquire a sample dataset, the sample dataset including at least one set of sample normalization values ​​and a sample consumption index corresponding to each set of sample normalization values; The sample weighting unit is used to obtain a preset training model and to perform weighting processing on the normalized values ​​of all the samples through the preset training model to obtain the predicted consumption index. The loss prediction unit is used to determine the predicted loss value of the preset training model based on the sample consumption index and the predicted consumption index corresponding to the same sample normalization value. The model convergence unit is used to iteratively update the initial parameters in the preset training model when the predicted loss value does not reach the preset convergence condition, until the predicted loss value reaches the convergence condition, and then record the converged preset training model as the preset fusion model.

[0083] A computer device includes a memory, a processor, and a computer program stored in the memory and executable on the processor, the processor being used to perform the above-described tool life prediction method based on multi-source data.

[0084] Specific limitations regarding the computer equipment, processor, and their various units and modules can be found in the above-described limitations of the tool life prediction method based on multi-source data, and will not be repeated here. Each module in the aforementioned processor can be implemented entirely or partially through software, hardware, or a combination thereof. Understandably, the processor includes a processor, memory, network interface, and database connected via a device bus. Each module of the processor can be embedded in hardware or independent of the processor, or stored in memory as software, so that the processor can call and execute the operations corresponding to each module. The processor provides computational and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores operating devices, computer programs, and a database. The internal memory provides an environment for the operation of the operating devices and computer programs in the non-volatile storage media. The database stores the data used in the tool life prediction method based on multi-source data in the above embodiments. The network interface is used for communication with external terminals via a network connection. When the computer program is executed by the processor, it implements a tool life prediction method based on multi-source data.

[0085] In one embodiment, a computer-readable storage medium is provided, the computer-readable storage medium storing a computer program that, when executed by a processor, implements the above-described tool life prediction method based on multi-source data.

[0086] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium. When executed, the computer program can include the processes of the embodiments of the above methods. Any references to memory, storage, databases, or other media used in the embodiments provided in this application can include non-volatile and / or volatile memory. Non-volatile memory may include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM), or flash memory. Volatile memory may include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in a variety of forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), dual data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous link DRAM (SLDRAM), RAMbus direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM), etc.

[0087] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the above-described division of functional units and modules is used as an example. In practical applications, the above functions can be assigned to different functional units and modules as needed, that is, the internal structure of the device can be divided into different functional units or modules to complete all or part of the functions described above.

[0088] The above-described embodiments are only used to illustrate the technical solutions of the present invention, and are not intended to limit it. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention, and should all be included within the protection scope of the present invention.

Claims

1. A tool life prediction method based on multi-source data, characterized in that, include: Acquire real-time data from multiple sensors and analyze the tool program sheet to determine theoretical data; The real-time data includes the actual cutting length and the actual spindle speed; The real-time data is normalized using the theoretical data to obtain parameter normalization values ​​corresponding to each of the real-time data. The material hardness coefficient is obtained, and the normalized values ​​of all the parameters and the material hardness coefficient are weighted by a preset fusion model to obtain tool consumption data. Determine the cutting deviation rate corresponding to the actual cutting length and the rotational speed deviation rate corresponding to the actual spindle speed, and determine whether the cutting deviation rate and the rotational speed deviation rate meet the preset correction conditions; When the cutting deviation rate and / or the rotational speed deviation rate meet the preset correction conditions, the tool consumption data is corrected using the cutting deviation rate and the rotational speed deviation rate to obtain corrected consumption data; A life prediction model is obtained, and the life prediction model is used to predict the life of the corrected consumption data to obtain the tool prediction life. Specifically, the tool consumption rate is calculated based on the tool consumption data of different time periods, the tool remaining index is calculated based on the tool consumption data, and the result of dividing the tool remaining index by the tool consumption rate is determined as the tool prediction life.

2. The tool life prediction method based on multi-source data as described in claim 1, characterized in that, The theoretical data includes theoretical cutting length and theoretical spindle speed; The determination of the cutting deviation rate corresponding to the actual cutting length and the speed deviation rate corresponding to the actual spindle speed includes: The absolute value of the cutting difference between the actual cutting length and the theoretical cutting length is determined, and the cutting deviation rate is determined based on the absolute value of the cutting and the theoretical cutting length. Determine the absolute value of the difference between the actual spindle speed and the theoretical spindle speed, and determine the speed deviation rate based on the absolute value of the speed and the theoretical spindle speed.

3. The tool life prediction method based on multi-source data as described in claim 1, characterized in that, The preset correction conditions include a preset cutting threshold and a preset rotation speed threshold; The step of determining whether the cutting deviation rate and the rotational speed deviation rate meet the preset correction conditions includes: The cutting deviation rate and the preset cutting threshold are compared to obtain the cutting comparison result. The speed deviation rate and the preset speed threshold are compared to obtain the speed comparison result. When the cutting comparison result indicates that the cutting deviation rate is greater than the preset cutting threshold, and / or the rotational speed comparison result indicates that the rotational speed deviation rate is greater than the preset rotational speed threshold, it is determined that the cutting deviation rate and the rotational speed deviation rate meet the preset correction condition; When the cutting comparison result indicates that the cutting deviation rate is less than or equal to the preset cutting threshold, and the rotational speed comparison result indicates that the rotational speed deviation rate is less than or equal to the preset rotational speed threshold, it is determined that the cutting deviation rate and the rotational speed deviation rate do not meet the preset correction conditions.

4. The tool life prediction method based on multi-source data as described in claim 1, characterized in that, The process of correcting the tool consumption data using the cutting deviation rate and the rotational speed deviation rate to obtain corrected consumption data includes: Obtain a preset correction coefficient, and determine a correction deviation value based on the preset correction coefficient, the cutting deviation rate, and the rotational speed deviation rate; The tool consumption data is corrected by using the correction deviation value to obtain corrected consumption data.

5. The tool life prediction method based on multi-source data as described in claim 1, characterized in that, The preset fusion model includes: Life_Index=C1T norm +C2L norm +C3S norm +C4F norm +C5M; in: Life_Index represents tool consumption data; C1, C2, C3, C4, and C5 are preset weighting coefficients; T norm This is a time normalization value; L norm This is the normalized value for cutting. S norm This is the normalized value for rotational speed; F norm This is the load normalization value; M is the material hardness coefficient of the cutting tool.

6. The tool life prediction method based on multi-source data as described in claim 1, characterized in that, After obtaining tool consumption data by weighting all the normalized values ​​of the parameters and the material hardness coefficient using a preset fusion model, the process further includes: Obtain all warning levels, and associate each warning level with a warning index range; According to the order of the warning levels from high to low, the tool consumption data is compared with the warning index range corresponding to each warning level in turn to determine the warning index range to which the tool consumption data belongs; The warning level corresponding to the warning index range to which the tool consumption data belongs is determined as the target warning level, and the triggering of the target warning level is determined.

7. The tool life prediction method based on multi-source data as described in claim 1, characterized in that, Before obtaining the material hardness coefficient, the method further includes: Obtain a sample dataset, which includes at least one set of sample normalization values ​​and a sample consumption index corresponding to each set of sample normalization values; A preset training model is obtained, and the normalized values ​​of all the samples are weighted using the preset training model to obtain the predicted consumption index. The prediction loss value of the preset training model is determined based on the sample consumption index and the prediction consumption index corresponding to the same sample normalization value. When the predicted loss value does not reach the preset convergence condition, the initial parameters in the preset training model are iteratively updated until the predicted loss value reaches the convergence condition. Then, the converged preset training model is recorded as the preset fusion model.

8. A tool life prediction device based on multi-source data, characterized in that, include: The data acquisition module is used to acquire real-time data measured by multiple sensors and to analyze the tool program sheet to determine theoretical data. The real-time data includes the actual cutting length and the actual spindle speed; The data normalization module is used to normalize the real-time data using the theoretical data to obtain parameter normalization values ​​corresponding to each of the real-time data. The consumption index module is used to obtain the material hardness coefficient. By using a preset fusion model, the normalized values ​​of all the parameters and the material hardness coefficient are weighted to obtain tool consumption data. The deviation rate determination module is used to determine the cutting deviation rate corresponding to the actual cutting length and the rotational speed deviation rate corresponding to the actual spindle speed, and to determine whether the cutting deviation rate and the rotational speed deviation rate meet the preset correction conditions. The data correction module is used to correct the tool consumption data by means of the cutting deviation rate and / or the speed deviation rate when the cutting deviation rate and / or the speed deviation rate meet the preset correction conditions, so as to obtain corrected consumption data. The tool life prediction module is used to acquire a tool life prediction model, and to predict the tool life based on the corrected consumption data using the tool life prediction model. Specifically, the tool consumption rate is calculated based on the tool consumption data of different time periods, the tool remaining index is calculated based on the tool consumption data, and the result of dividing the tool remaining index by the tool consumption rate is determined as the tool life prediction.

9. A computer device, characterized in that, It includes a memory, a processor, and a computer program stored in the memory and executable on the processor, the processor being used to perform the tool life prediction method based on multi-source data as described in any one of claims 1 to 7.

10. A computer-readable storage medium storing a computer program, characterized in that, When the computer program is executed by the processor, it implements the tool life prediction method based on multi-source data as described in any one of claims 1 to 7.

Images